Australian researchers have developed a robot capable of distinguishing crown-of-thorns starfish in coral reef ecosystems and killing the creature with a single shot of salt.
The robot, equipped with state-of-the-art computer vision and a machine-learning system, could play an important role in protecting Australia's Great Barrier Reef from the creature, which is believed to be responsible for up to 40 per cent of the coral cover decline across the reef.
The crown-of-thorns sea star (COTS), with its typical crown of venomous thorn-like spikes covering its upper surface, is one of the largest sea stars. It preys on coral polyps. A single starfish can consume up to six square metres of living coral reef per year.
Conservationists fight against the species by sending human divers into COTS hotspots to inject the animals with bile salt.
The robot, dubbed COTSbot, developed by a team from Queensland University of Technology in Brisbane, Australia, could simplify and speed up the process.
"We see the COTSbot as a first responder for ongoing eradication programmes, deployed to eliminate the bulk of COTS in any area, with divers following a few days later to hit the remaining COTS,” explained the robot’s creator Matthew Dunbabin.
"The COTSbot becomes a real force multiplier for the eradication process the more of them you deploy. Imagine how much ground the programs could cover with a fleet of 10 or 100 COTSbots at their disposal, robots that can work day and night and in any weather condition."
The robot, equipped with stereoscopic cameras for in-depth perception, can stay in water for up to eight hours and kill more than 200 COTSs.
The researchers tested the robot, believed to be the first of its kind in the world, in a trial in Queensland’s Moreton Bay and hope to deploy it at the Great Barrier Reef in December this year.
In the latest trials, they focused on testing the robot’s mechanical parts and navigation system. The robot uses five thrusters to maintain stability and is equipped with GPS and pitch-and-roll sensors.
The researchers spent six week developing and training the robot to recognise COTS among coral, feeding thousands of still images and videos of the reef into its software.
"Its computer system is backed by some serious computational power, so COTSbot can think for itself in the water," said Feras Dayoub, who designed the COTS-detecting software.
"If the robot is unsure that something is actually a COTS, it takes a photo of the object to be later verified by a human and that human feedback is incorporated into the robot's memory bank.”
The researcher explained that making sure the robot can reliably recognise COTS was not an easy feat as the coral reef environment is very complex. The robot’s camera and algorithms have to be able to overcome low visibility as well as depth-dependent colour changes.
The researchers first stumbled upon the idea to use a robot to fight COTS some ten years ago. However, back then, it required up to 20 injections to kill a single starfish. It was only last year, when researchers from James Cook University (JCU) developed a new single-shot method, that the team was able to bring the idea seriously back to life.
"That was the game-changer that opened the doors for a robotic solution to the COTS problem,” Dunbabin said. “Combining this with new advances in machine learning meant we could make COTSbot a reality."
The team believes COTSbot is the first autonomous underwater vehicle to be equipped with an injection system. It is also designed to operate exclusively within a metre of the seafloor, which is extremely challenging for any robot.
The researchers will take the robot for first tests at the Great Barrier Reef next month to test its performance recognising the starfish. During the trial, every action of the robot will be first verified by a human controller to prevent mistakes.
The Great Barrier Reef's COTS starfish-killing robot infographic